As the old Danish proverb says, it is difficult to make predictions, especially about the future. Add the exponential pattern of change familiar from Moore’s Law to the mix, multiply over a combination of dynamic trends, and such augury is fraught with peril.
Take, for example, the Internet of Things (IoT). Simply looking at the devices (sensors and actuators) themselves, we may extrapolate a number of such trends, many of which follow exponential curves.
Such devices are rapidly getting smaller. They’re consuming less and less power, so little that many such devices never require a battery replacement, if they have a battery at all. Simultaneously, the amount of processing power we can place on such tiny devices is also exploding.
Then there’s the somewhat discontinuous improvements in connectivity, as we move ever closer to – and finally take the leap to – 5G. And last but certainly not least, the cost of each device is also dropping to the point where we no longer will even need to think about the cost per unit, but rather the cost per thousands or even millions of units.
Today, we think of IoT devices in discrete terms: home automation products or factory devices or smart city sensors and the like.
Tomorrow, however, all the various disruptive trends impacting the IoT will combine to completely change how we perceive of the IoT and the uses we might wish to apply it toward. Here is a glimpse.
If we focus our attention solely on sensors for a moment, the trend is moving away from individually installed and configured sensors to fabrics of sensors – often literally. Manufacturers of everything from textiles to building materials to factory equipment will simply include sensors in their manufacturing processes, essentially putting them everywhere.
Imagine if you will, a retail floor so densely packed with sensors that they can pick up the movements of insects scurrying across a store aisle. Or a component of a piece of factory equipment so well-instrumented that its digital twin provides resolution down to the micrometer.
Today, connectivity is a limiting factor for such a vision, as wired connectivity would soon become impractical and Wi-fi is certainly not up to the task. However, 5G allows for far greater ‘densification’ of endpoints (that is, how many you can pack into a given volume) than ever before.
Yes, in our vision of IoT sensor fabric, every sensor is potentially its own 5G endpoint – its own self-contained cellphone as it were.
The Big Data challenges – and opportunities – from such fabric also boggle the mind. At some point, sensors will be so close together that neighboring ones will often deliver the same data. In other words, data from IoT sensor fabrics will have much higher redundancy than today’s IoT applications.
Your first reaction might be that such redundant data is useless data – but in fact, it serves a few vital purposes. First, it serves as a calibration check. When you have millions of sensors, some of them will be miscalibrated or otherwise giving you false readings.
If you only have a handful of such sensors, then it’s difficult to single out the bad ones. But if you have thousands of redundant sensors, the anomalous data from the bad ones will stand out like sore thumbs.
As a result, you’ll never want to bother fixing or replacing miscalibrated sensors. You’ll simply ignore their telemetry – regardless of whether their errors are accidental or malicious.
Malicious miscalibrations – what we call calibration attacks – are a little known, but dangerous attack vector for the IoT (the VW emissions scandal from a few years back is the best-known example. See my Cortex from the time for a discussion).
However, while tampering with a small number of sensors may give an attacker a means for achieving their goals, vast numbers of sensors will be prohibitively difficult to compromise.
Sensors require little or no processing power, as their role is simply to collect and then deliver data upstream. Actuators, in contrast, have a reason to leverage their own processors, as they must take action.
Today’s basic actuators may do little more than respond to control signals from elsewhere, but as processors get smaller, cheaper, and more powerful, we’ll be seeing actuators become increasingly intelligent.
In fact, the word ‘actuator’ is a mechanical term that denotes motion, as in triggering a switch on a piece of equipment. However, if we couple the actuator with a transceiver (that is, with bidirectional connectivity), then the action it takes may simply be to send a signal – either upstream to a gateway or controller, or to other IoT devices.
When we envision smart IoT devices being able to communicate with each other in large numbers, a fascinating development results: emergent behaviors.
Emergent behaviors are behaviors of complex systems of systems (in our case, IoT devices) that only appear at the aggregate level, but not at the level of the devices (see my discussion of emergence from 2015).
In other words, the coded behavior in each device might be quite simple, but the behavior of large numbers of communicating devices might be entirely different from the original intentions of the coders.
In some circles, people call such behavior swarm behavior – although the word swarm is used in similar, but different ways that can be needlessly confusing, like the notions of a Docker Swarm or an Ethereum Swarm.
Emergent behaviors are quite common in the natural world, and explain much of its complexity. For example, bees in a beehive follow simple, instinctive behaviors individually, while collectively they can build the honeycomb structure of the hive.
The problem we face today is that the techniques for programming IoT actuators so that the resulting emergent behaviors are desirable are largely unknown, or at best experimental. That doesn’t mean, however, that we should throw in the towel.
On the contrary, such behaviors will emerge whether we want them or not. If we don’t get a handle on how to direct them in useful ways, we’ll end up with alarmingly bad results.
On the plus side, our best bet for providing adequate comprehensive security for the IoT is via programming the right emergent behaviors. IoT swarms must behave as though they have their own immune systems, identifying and counteracting malicious behavior regardless of the form it takes.
Our only hope of getting a handle on such security, therefore, is to figure out how to program such emergent immune system behavior.
There’s one important trend I have yet to mention: blockchain. If you attempt to apply ‘true’ blockchain to the IoT, however, you quickly run into the challenges of how to handle blockchain nodes within the context of an IoT fabric.
If you extend your view beyond what is already being thought of as ‘first generation’ blockchain to the broader application of distributed ledger technologies, then the light at the end of this tunnel suddenly brightens.
One particularly promising effort: IOTA. IOTA bills itself as a ‘blockless distributed ledger,’ and is thus technically not a blockchain-based technology at all.
While blockchain-based systems require a consensus of transaction processing nodes (aka ‘miners’) to complete transactions, IOTA drives consensus into the peer-to-peer endpoints themselves – no miners needed.
The IOTA architecture, therefore, is well-suited to the IoT fabric this article discusses – at least in theory. IOTA is still a work in progress, and today the effort is focused on transactions of value (for example, one device ‘paying’ another device to take an action).
Such transactionality will undoubtedly be a part of the IoT in years to come, but in my opinion, will be but one use case among many.
Will IOTA or some other post-blockchain innovation drive the dominant architecture of next-generation IoT? Your guess is as good as mine.
Copyright © Intellyx LLC. Intellyx publishes the Agile Digital Transformation Roadmap poster, advises companies on their digital transformation initiatives, and helps vendors communicate their agility stories. As of the time of writing, none of the organizations mentioned in this article are Intellyx customers. Image credit: Emma Jane Hogbin Westby.